Chapter 7 Process-related problems

241

Chapter 7 Process-related problems

7.1 Introduction

The objective of this chapter is to address important process-related problems. Most of these problems are the result of our skewed emphasis on turbidity currents and their deposits. These problems have manifested themselves into popular turbidite myths. Ten of these myths have been dispelled (Shanmugam, 2002a). 7.2 Conflicting definitions of turbidity currents

Kuenen (1951) used the term 'turbidity currents of high-density' for stratified flows with an upper turbulent layer and a lower non-turbulent (slide/debris flow) layer (Fig. 7.1A). In justifying his definition, Kuenen (see comment in Sanders, 1965, p. 217) explained that '/t is much simpler to leave out of the definition of turbidity current any reference to hydrodynamic mechanisms.' However, one cannot practice process sedimentology without using basic principles of hydrodynamics. Because of this loophole in Kuenen's definition, later authors have misused the concept of turbidity currents. Examples are" (1) McCave and Jones (1988, p. 250) advocated, '... deposition of ungraded muds from high-density non-turbulent turbidity currents.' (2) Kneller and Buckee (2000) claimed that turbidity currents can be non turbulent (i.e., laminar) in state. Furthermore, they claimed that turbidity currents are natural phenomena whose exact hydrodynamic properties are unclear. (3) Mutti et al. (2003b, p. 745) justified Kuenen's flawed definition for the sake of maintaining a stable terminology. Sanders (1965) was the first process sedimentologist who made a clear distinction between laminar debris flows and turbulent turbidity currents. He clarified that turbidity currents are density currents caused by sediment in turbulent suspension. This clarity comes from experimental results of Bagnold (1954, 1956)

242

Deep-waterprocesses andfacies models

Density-stratified flows

Fig. 7.1. (,4) Diagram showing Kuenen's (1950a, 1951) concept of'high-density turbidity currents' that includes both the basal non-turbulent (slide/debris flow) layer and the upper turbulent (turbidity current) layer. (B) Diagram showing Bagnold's (1956) concept that distinguishes basal laminar inertia region and upper turbulent viscous region as two different processes.

who distinguished turbulent flows from laminar flows in density-stratified flows (Fig. 7.1B). Middleton (1993, p. 93) clearly excluded laminar flows from turbidity currents by stating, '... if it is truly nonturbulent it can no longer be classified as a turbidity current.' Thus the characterization 'non-turbulent turbidity currents' is an oxymoron from a hydrodynamic point of view. 7.3 C o n f l i c t i n g d e f i n i t i o n s of t u r b i d i t e s

The crux of the turbidite controversy can be attributed to conflicting definitions of turbidity currents and turbidites. For example: (1) Kuenen (see comment in Sanders, 1965, p. 218) defined, 'Deposits from all kinds and combinations o f currents falling under the definition o f turbidity currents are turbidites, whether there was bottom traction, laminar flow, non-turbulent flow etc., involved or not.' (2) Following the approach of Kuenen, Mutti (1992, p. 40), stated, 'Cohesive debris flows and turbidity currents should therefore be considered the two main mechanisms responsible for having transported and deposited the bulk o f turbidite sediments.'

Process-related problems

243

(3) Furthermore, Mutti et al. (1999, p. 19) defined 'turbidites' as the deposits of all sediment-gravity flows, which include debris flows, grain flows, fluidized sediment flows, and turbidity currents (Fig. 7.2). This approach of Mutti et al., if applied, would undo the progress that have been made in process sedimentology during the past four decades in distinguishing deposits of debris flows, grain flows, fluidized/liquefied flows, and turbidity currents from one another. (4) Sanders (1965) emphasized that the term turbidites should refer strictly to those deposits that formed from turbulent suspension of turbidity currents (Fig. 7.2). Middleton and Hampton (1973) also considered only those deposits of turbidity currents to be turbidites. In spite of this simple and straight forward concept, the term turbidite means different things to different researchers. To some, turbidite means any deep-water sand, to others, turbidite means a deep-water channel or lobe sand, but to process sedimentologists turbidite means deposit of a turbidity current solely. (5) Kneller (1996, p. 76) claimed, '... many features recently suggested as characteristic of debris flow (e.g. sharp upper grain-size breaks, floating or

Sediment- Gravity Flows (Middleton and Hampton, 1973) |

Turbidity Currents .

.

.

.

.

Fluidized Flows

Debris Flows

Grain Flows

.

Turbidit (Sanders, 1965) Turbidites (Mufti et al., 1999) Fig. 7.2. Two differing definitions of the term 'turbidites" According to Sanders (1965), turbidites are the exclusive deposits of turbidity currents. According to Mutti et al. (1999), deposits of all sediment-gravity flows, which include turbidity currents, fluidized flows, debris flows, and grain flows, are turbidites. In this book, Sanders' (1965) definition is followed. (After Shanmugam (2002a). Reproduced with permission from Elsevier.)

244

Deep-water processes and facies models

rafted mudstone clasts, inverse grading of clasts, moderate to high matrix content) or contourites (e.g. traction structures) are to be expected in turbidites.' In other words, Kneller would interpret features that are characteristic of debrites and contourites as that of turbidites. Because of these conflicting definitions of turbidites, deposits of debris flows and avalanches have been classified as 'turbidites' using exotic nomenclature. Selected examples with my comments are (see Table 2.2)" (1) (2) (3) (4)

Fluxoturbidites (Dzulynski et al., 1959): deposits of sand avalanches. Seismoturbidites (Mutti et al., 1984): deposits of large-scale mass flows. Megaturbidites (Labaume et al., 1987): deposits of debris flows. Atypical turbidites and problematica (Stanley et al., 1978): deposits of slumps, debris flows, and sand flows. (5) High-concentration sandy turbidites (Abreu et al., 2003): deposits of sandy debris flows.

In explaining the popularity of turbidites, Carter (1975, p. 147) wrote that '... the temptation is always to tailor field observations to presently known processes of sediment deposition, rather than to tie them to speculative theoretical possibilities; it is therefore not surprising that many published studies of flysch sequences place great emphasis on features explicable by turbidity current hypothesis, and tend to be somewhat skeptical regarding deposition of individual beds by other mass-transport processes.' Lowe (1979, p. 81) observed that '/t is too often convenient to interpret deposits exhibiting textures and structures whose origins are problematic in terms of processes whose dynamics are equally poorly understood.' Fortunately, significant process sedimentological advances have been made during the past 50 years based on theoretical (Bagnold, 1954, Dott, 1963; Sanders, 1965, Middleton and Hampton, 1973; Enos, 1977; Allen, 1985a; Shanmugam, 1996a), experimental (Hampton, 1972; Marr et al., 2001), and observational (Hollister, 1967; Fisher, 1971; Shepard et al., 1979; Shanmugam et al., 1993a; Stow et al., 1998) basis. Thus we can now differentiate deposits of turbidity currents from those of other processes with confidence (see Chapters 3, 4, and 5).

7.4 Conflicting definitions of high-density turbidity currents The concept of 'high-density turbidity currents' has been the center of controversy for more than 50 years. This is because the concept has been defined on the basis of four conflicting properties, namely, (1) flow density (Kuenen, 1950a) or grain concentration (Pickering et al., 2001); (2) driving force (Postma et al., 1988); (3) grain size (Lowe, 1982); and (4) flow velocity (Kneller, 1995). These definitions

Process-related problems

245

are inconsistent with one another in terms of Newtonian rheology, turbulent state, and principal sediment-support mechanism of turbulence, the properties that define turbidity currents (Dott, 1963; Sanders, 1965; Middleton, 1993). As a consequence, virtually any process can be classified as a high-density turbidity current. 7.4.1 Flow density

Based on his experiments with subaqueous mud flows, Kuenen (1950a, p. 44) envisioned the concept of high-density turbidity currents. Kuenen conducted three series of experiments using an aquarium, a ditch, and a tank. In his experiments using an aquarium of 2 m in length and 50 cm in depth and breadth, he used slurries of clay, sand, and gravel with flow densities of up to 2 g/cm 3 on a slope of 8.5 ~ Unfortunately, Kuenen used the wrong term, 'turbidity currents of high density,' for density-stratified flows with densities of 2 g/cm 3. In terms of flow densities (e.g., Hampton, 1972), Kuenen's experimental flows are clearly debris flows. The clay content (23 to 33% of total solids by weight) in Kuenen's (1951) experiments was so high that his experimental flows were considered to be debris flows (Oakeshott, 1989). In Kuenen's (1951, p. 15) experiments, flows were observed to slide down a slope, move like a glacier, break up into slabs, crack on their surface, and come to rest on the slope (Fig. 7.3). These physical behaviors are typical of slides and debris flows rather than those of turbidity currents (see Chapter 3).

Fig. 7.3. Conceptual diagram illustrating properties of Kuenen's (1951) experimental 'Turbidity currents of High Density" Note most properties are suggestive of deposition from debris flow, not turbidity currents.

246

Deep-water processes and facies models

In spite of its emphasis on flow density, the concept of high-density turbidity current has never been defined consistently on the basis of flow density (Shanmugam, 1996a). For example: (1) The distinction between 'low' and 'high' density currents was set at a density value of 1.1 g/cm 3 (Kuenen, 1966). (2) High-concentration flows (i.e., high-density turbidity currents) have a density range of 1.5 to 2.4 g/cm 3 (Middleton and Hampton, 1973). (3) A debris flow has a density of 2.0 g/cm 3 (Hampton, 1972). Thus high-density turbidity currents and debris flows are one and the same from a density point of view (Fig. 7.4A). Different researchers have used different sediment concentration values for high-density turbidity currents (Table 7.1 and Fig. 7.4B). For example: (1) Kuenen's (1966) density value of 1.1 g/cm 3 converts into 6% concentration of solids by volume (Picketing et al., 1989, p. 17). This means that high-density behavior can begin at a low concentration value of 6% by volume. (2) Middleton (1967) used a concentration value of 44% by volume in his experiments of high-density flows. (3) Lowe (1982) considered that the sediment concentration must be greater than 20-30% by volume for the onset of high-density behavior. (4) Postma et al. (1988) used a concentration value of 35-40% by volume in their experiments of high-density flows. (5) Leclair and Arnott (2005) used concentration values of 20, 25, and 35% by volume in their experiments of high-density flows. (6) Subaerial hyperconcentrated flows are often compared with subaqueous high-density turbidity currents. According to Pierson and Costa (1987), hyperconcentrated flow, which is intermediate between stream flow and debris flow (Beverage and Culbertson, 1964), has a sediment concentration of 20-60% by volume (Fig. 7.4B). Furthermore, Qian et al. (1980) considered a Bingham fluid (i.e., a debris flow) as a special homogeneous type of hyperconcentrated flow. Because of these wide ranges of concentration values, a flow with 20% sediment concentration would be considered a low-density flow by Middleton, a high-density flow by Kuenen, a hyperconcentrated flow by Pierson and Costa, and a debris flow by Qian et al. In order to differentiate high-density turbidity currents from low-density turbidity currents, there must be a defining concentration or density value. But no such defining density value has been established. A fundamental issue here is the upper limit of sediment concentration that controls fluid turbulence in Newtonian turbidity currents. Bagnold (1954, 1956) investigated aggregates of cohesionless grains in Newtonian fluid under shear

Process-related problems

247

Fig. 7.4. (A) Plot of flow density for different flow types. Note overlapping density values between high-concentration flows (i.e., high-density turbidity currents) and debris flows. (B) Range of sediment concentration for different flow types. Note a flow with 20% sediment concentration can be classified as any one of the three types. (After Shanmugam (1996a). Reproduced with permission from SEPM.)

and introduced the concepts of 'inertia' and 'viscous' regions based on grain inertia and fluid viscosity, respectively (Fig. 7.1B). In a grain-inertia region the effects of grain concentration and related grain collision (i.e., dispersive pressure) dominate, whereas in a viscous region effects of fluid viscosity dominate. More importantly, in the grain-inertia region high grain concentration tends to suppress

Deep-water processes and facies models

248

T a b l e 7.1 V a r i a b l e concentration values of different flow types

Low-density turbidity current 9 < 1% sediment by weight (Middleton, 1993) 9 < 6% solids by volume (Picketing et al., 1989) or < 1.1 g/cm 3 (Kuenen, 9 23% solids by volume (Middleton, 1967)

1966)

Itigh-density turbidity current 9 9 9 9 9 9

> 6% solids by volume (Picketing et al., 1989) or > 1.1 g/cm 3 (Kuenen, 1966) > 10% sediment by weight (Middleton, 1970, 1993) > 20-30% solids by volume (Lowe, 1982) 20, 25, and 30% sediment by volume (Leclair and Amott, 2005) 35-40% solids by volume (Postma et al., 1988) 44% solids by volume (Middleton, 1967)

Hyperconcentrated stream flow 9 20-60% sediment by volume or 40-80% by weight (Pierson and Costa, 1987)

fluid turbulence and promote laminar shear. In explaining suppression of turbulence by grain concentration I, Bagnold (1956, p. 288) stated, 'At a certain stage the turbulence began to be suppressed, being damped out by the increasing overall shear resistance. And on a further increase in the grain population the turbulence vanished altogether .... When the turbulence finally ceased, C was apparently uniform from top to bottom, at about 0.3 .... Having attained uniformity in the now laminar fluid flow, C could be increased nearly to the mobile limit of about 0.53. Ultimately the whole flow froze'simultaneously at all depths.'

Because a turbulent flow becomes laminar when its grain concentration reaches 30%, turbidity currents cannot theoretically exist at these high concentration values. Bagnold (1962) later demonstrated that gravity-driven turbulent currents could be maintained only in flows with low grain concentration of less than 9% by volume. To date, no one has ever demonstrated in flume experiments that fully turbulent flows (i.e., flows without basal high-concentration laminar layer) can be created and maintained using sediment concentration in excess of 9% by volume. Neither has any one documented such flows in the modem deep sea. 7.4.2 Driving force

Postma et al. (1988) used the concept of high-density turbidity current for densitystratified flows in which the basal high-concentration laminar layer is derived from and driven by over-tiding turbidity currents (Fig. 7.5). However, the basal high-concentration layer (i.e., sandy debris flow) cannot be a turbidity current because of its pseudoplastic rheology and its laminar flow state (Fig. 7.5). The basal

Process-relatedproblems

249

Newtonian (Fluidal)

Turbulent High-Density

Turbidity Current

I:i: ./.':".'.~.~.'..'_:~.i'i."_i'.i~'::~~ [~eb/is Flow '-" " A. Flow Type (This Study)

NonLaminar Newtonian (Pseudoplastic) B. Behavior (Rheology)

C. State

D. Conventional

Usage

(Postma et al, 1988)

Fig. 7.5. (A) Side view of experimental density-stratified flows. Postma et al. (1988) suggested that the basal high-concentration layer (labeled as sandy debris flow in this book) was driven by the upper low-concentration layer (labeled as turbidity current). Note mudstone clasts at the upper rheological boundary of sandy debris flow. (B) Lower (non-Newtonian) and upper (Newtonian) layers represent two separate rheologies. (C) Lower (laminar) and upper (turbulent) layers represent two different flow states. (D) Conventional usage has been to lump both lower and upper layers together under the term 'high-density turbidity currents" (From Shanmugam (1997a). Reproduced with permission from Elsevier.)

layer with high sediment concentration would severely hinder settling, and would freeze mudstone clasts in floating positions along a rheological boundary (Fig. 7.5). The basal laminar and upper turbulent layers in high-density turbidity currents are analogous to 'inertia' and 'viscous' regions of Bagnold (Fig. 7.1B). In justifying their use of the concept of high-density turbidity currents, Mutti et al. (1999, p. 22) stated, 'The concept of basal granular layer coincides with that of 'high-density turbidity current' (cf Lowe, 1982), the latter term being very popular among many sedimentologists and probably the easiest to use for general purposes of communication.'

Such a practice is called popular sedimentology, not process sedimentology. We should use a concept because it is hydrodynamically sound, not because it is popular. Other authors have also expressed doubts about the concept of highdensity turbidity currents. (1) Hallworth and Huppert (1998, p. 1083), based on their experiments on highconcentration gravity flows with density stratification, stated that '... we are still unsure o f the p h y s i c a l causes b e h i n d the effects we p r e s e n t here . . . . '

250

Deep-water processes and facies models

(2) Kneller and Buckee (2000, p. 87) emphasized that '... existing theory seems inadequate to explain the behavior of some highly mobile dense dispersions, and arguments based solely on the geological interpretation of deposits may be inadequate to resolve issues of process.' (3) Pickering and Hilton (1998, p. 89) concluded that '... the precise hydrodynamic conditions and sediment concentrations of high-concentration turbidity currents remains unresolved.' Disappointingly, these authors went on to apply the concept of high-density turbidity currents in their subsequent studies (e.g., Picketing et al., 2001).

Fig. 7.6. An illustration of Lowe's (1982) classification of turbidity currents based on grainsize populations. Population 1 = low-density turbidity currents; population 2 = sandy highdensity turbidity currents; population 3 = gravelly high-density turbidity currents. Rheology, depositional mechanism, and flow type are added for clarification. Because hindered settling, matrix buoyant lift and dispersive pressure are important sediment support mechanisms in 'high-density turbidity currents" High-density turbidity currents, as defined by Lowe (1982), are considered to be more like plastic debris flows than Newtonian turbidity currents. (After Shanmugam and Moiola (1995). Reprinted by permission of the American Association of Petroleum Geologists whose permission is required for further use.)

Process-related problems

251

7.4.3 Grain size

Lowe (1982) classified turbidity currents into two principal types, namely lowdensity flows and high-density flows. He based his classification on grain size populations, particle concentrations, and sediment support mechanisms (Fig. 7.6). Low-density turbidity currents are composed of population 1 grains (clay to medium-grained sand) in which the sediment-support mechanism (i.e., turbulence) is independent of particle concentration. As defined by Lowe (1982), these flows in terms of fluid rheology and sediment-support mechanism (i.e., flow turbulence) are true turbidity currents. In high-density turbidity currents, sediment support is concentration dependent (above 20-30%). In sandy high-density flows composed of population 2 (coarse sand to small pebble), grains are supported by hindered settling and turbulence. In gravelly high-density flows containing population 3 (pebble and cobble), grains are supported mainly by matrix buoyant lift, and dispersive pressure (Lowe, 1982, p. 283). Because hindered settling, matrix buoyant lift, and dispersive pressure are the predominant sediment-support mechanisms in highdensity turbidity currents (Lowe, 1982, p. 282), such high-density flows are more akin to plastic debris flows than they are to Newtonian turbidity currents. This conclusion validates Bagnold's (1962) original observation that true turbidity currents are low-density flows. And these low-density turbidity currents are incapable of carrying coarse sand and gravel in turbulence. Thus there is no need to classify turbidity currents into low- and high-density types. 7.4.4 Flow

velocity

Kneller (1995) realized that simple normally graded beds are rare in deep-water sequences, and that most deep-water sequences with complications (e.g., massive sandstones, disordered sequences, abrupt grain-size breaks, large-scale bedforms, etc.) are difficult to interpret as turbidites. To alleviate this problem, Kneller has redefined turbidity currems using velocity (u), distance (x), and time (t). He classified turbidity currents into five types: (1) depletive waning flow; (2) uniform waning flow; (3) depletive steady flow; (4) depletive waxing flow; and (5) accumulative waning flow. This classification allows room for interpreting deepwater deposits with any kind of grading (i.e., normal or inverse) as turbidites (Fig. 7.7). Thus the very foundation of the turbidite paradigm, based on the relationship between turbidites and normal grading (Kuenen and Migliorini, 1950; Bouma, 1962), has been undermined by this redefinition. Unlike the conventional definition of turbidity currents, which is based on fluid rheology and flow turbulence, and that take into consideration sediment concentration, velocity, thickness, and viscosity, Kneller's (1995) definition is primarily concerned with velocity. He considered waxing flows as analogous to medium- to high-density turbidity currents (his p. 37). Redefinition of turbidity

252

Deep-water processes and facies models

Fig. 7.7. Three publications showing how opinions on nature of grading in turbidites have changed through time. Top: Bouma (1962) suggested normal grading for turbidites. Middle: Harms and Fahnestock (1965) proposed normal grading and massive (i.e., no grading) types for turbidites. Bottom: Kneller (1995) advocated normal grading, massive (i.e., no grading), and inverse grading for turbidites. (After Shanmugam (2000a). Reproduced with permission from Elsevier.)

currents using velocity alone has opened up new avenues for interpreting virtually all deep-water sands as turbidites despite whether they were deposited by turbidity currents or not. Rarity of normally graded beds in the rock record simply means that turbidites are rare in nature. This does not warrant a redefinition of turbidity currents. 7.4.5 Synonyms The untenable concept of high-density turbidity currents has resulted in a plethora of process terms for density-stratified flows. Gani (2004), for example, proposed the term dense flow for rheologically stratified flows that include both Newtonian (i.e., turbidity currents) and non-Newtonian fluids (i.e., debris flows). According to Gani (2004, his Fig. 3), the term dense flow is an alternative

Process-related problems

253

for: (1) high-density turbidity currents (Lowe, 1982); (2) sandy debris flows (Shanmugam, 1996a); (3) slurry flows (Lowe and Guy, 2000); (4) concentrated density flows (Mulder and Alexander, 2001); and (5) liquefied flows/fluidized flows. When viewed in isolation, Gani's approach appears to be reasonable. However, when viewed in total from an historical perspective, a troubling trend emerges: (1) The term dense flow was originally applied by Norem et al. (1990, Fig. 3 therein) strictly to the basal (debris flow) layer in density-stratified flows. This is meaningful because debris flows having higher densities than turbidity currents are bound to occur at the base. Gani (2004), however, applied the term dense flow for both basal (debris flows) and upper (turbidity currents) layers. Gani borrowed the concept of dense flow for a hybrid of turbidity current and non-cohesive debris flow from Allen (1997), who, in turn, borrowed it from Mulder and Cochonat (1996). Surprisingly, neither Mulder and Cochonat nor Gani acknowledged the conflict with the earlier use of the term by Norem et al. (2) The term gravitite was originally introduced by Natland (1967) to represent deposits of primarily debris flow. Without acknowledging the contribution of Natland, Gani (2004) coined a slightly different term, gravite, for deposits of slide, slump, debris flow, dense flow, and turbidity current. (3) The term slurryflow was previously applied by Carter (1975) to the basal (debris flow) layer in density-stratified flows. Later, Lowe and Guy (2000) applied the term for both basal and upper layers. (4) The term inertia region, representing grain flow, was previously applied to the basal layer of density-stratified flows by Bagnold (1954, 1956). In contrast, Carter (1975) applied the term grain flow to the upper layer. According to Poliakov (2002), the term granular flow is an alternative for grain flow as well as for slumps, debris flows, high-density turbidity currents, and bottom avalanches. Various factors, which include failure to acknowledge original contributions, have resulted in a proliferation of 34 published synonyms for high-density turbidity currents. In density-stratified flows, these names represent either the basal laminar layer alone or both the basal laminar and the upper turbulent layer together (Table 7.2). These synonyms, compiled from published sources (Kuenen, 1950a, 1951; Bagnold, 1954,1956; Dzulynski et al., 1959; Dzulynski and Sanders, 1962; Sanders, 1965, 1981; Middleton, 1967; Carter, 1975; Friedman and Sanders, 1978; Fisher and Schmincke, 1984; Valentine, 1987; Postma et al., 1988; Norem et al., 1990; Friedman et al., 1992; Fisher, 1995; Shanmugam, 1996a, 2002a; Sanders and Friedman, 1997; Vrolijk and Southard, 1997; Mutti et al., 1999;

254

Deep-water processes and facies models

Table 7.2 Synonyms for density-stratified flows (i.e., high-density turbidity currents). Modified after Shanmugam (2002a) References *Bagnold (1954, 1956) Dzulynski and Sanders (1962) Sanders ( 1965)

Friedman and Sanders (1978), Sanders (1981), Friedman, Sanders and Kopaska-Merkel (1992, p. 335)

Sanders and Friedman (1997) *Kuenen (1951) Dzulynski et al. (1959) Carter (1975) *Postma et al. (1988) Norem et al. (1990, their Fig. 2) Shanmugam (1996a) *Marr et al. (2001) *Vrolijk and Southard (1997) *Kuenen (1950, 1951) *Middleton (1967) Lowe (1982) *Postma et al. (1988) Norem et al. (1990) Mutti et al. (1999) Lowe and Guy (2000) Mulder and Alexander (2001 ) McCaffrey et al. (2001) *Poliakov (2002) Gani (2004)

Upper layer (low-density) Viscous region Turbidity current

Lower layer (high-density) Inertia region Traction carpet

Turbidity current

Synonymous terms: Flowing-grain layer (p. 192) Inertia flow (p. 194) Fluidized flowing-grain layer (p. 210) Inertia-flow layer (p. 211) Avalanching flow (p. 213) Turbidity current Liquefied cohesionless-particle flow (Suspended load) ('Bed load') Synonymous terms: Inertial flow Grain flow Mass flow Rheologic bed stage Fluidized cohesionless-particle flow Turbidity current Liquefied cohesionless coarse-particle flow Turbidity current Slide Turbidity current Fluxoturbidity current Grain flow Slurry flow Turbulent suspension Laminar inertia-flow Suspension flow Dense flow (i.e., debris flow) Turbidity current Sandy debris flow Turbidity current Sandy debris flow Turbulent flow Laminar sheared layer Turbidity currents of high density High-concentration turbidity current High-density turbidity current High-density turbidity current Flowslide High-density turbidity current Slurry flow Concentrated density flow Particulate gravity currents Granular flow Dense flow

*Experimental studies.

Lowe and Guy, 2000; Marr et al., 2001; Mulder and Alexander, 2001; McCaffrey et al., 2001; Poliakov, 2002; Mulder et al., 2003; Gani, 2004), are" (1) Inertia flow (2) Inertia-flow layer (3) Inertial flow

Process-related problems

255

(4) Laminar-inertia flow (5) Laminar sheared layer (6) Traction carpet (7) Flowing-grain layer (8) Fluidized flowing-grain layer (9) Fluidized cohesionless-particle flow (10) Liquefied cohesionless-particle flow (11) Liquefied cohesionless coarse-particle flow (12) Avalanching flow (13) Slide (14) Slump (15) Mass flow (16) Grain flow (17) Debris flow (18) Sandy debris flow (19) Fluxoturbidity current (20) Pseudo-plastic quick bed (21) Bed load (22) Rheologic bed stage (23) Bottom avalanches (24) Stratified flow (25) Density-stratified flow (26) Nude ardente (i.e., applied for decoupling of pyroclastic flows) (27) Slurry flow (28) High-concentration turbidity current (29) Flowslide (30) Concentrated density flow (31) Granular flow (32) Particulate gravity currents (33) Surge-type turbidity currents (34) Dense flow. This multiplicity of names may be a reflection of different researchers with different backgrounds (e.g., experimental versus outcrop) expressing their different perspectives. On the other hand, one might argue that inventing different names for density-stratified flows is a consequence (intended or unintended) of doing research on high-density turbidity currents repeatedly without addressing the fundamental problems. Such a practice might be construed as an act of insanity, as defined by Albert Einstein: 'Doing the same thing over and over again and expecting different results' (Brainy Quote, 2005). The solution to this chronic problem is to define a flow layer based on its fluid rheology and flow state, irrespective of whether a flow is stratified or unified. Otherwise, the current trend would result in nearly 100 synonyms for high-density

256

Deep-water processes and facies models

turbidity currents; a jungle of jargon that sedimentary geologists would dread by the end of the 21st century! 7.5 Unknowable flow transformations

The transformation of one type of flow (e.g., laminar debris flow) into another (e.g., turbulent turbidity current) during transport is probably the single most important and the least understood phenomenon in deep-water process sedimentology. Flow transformations can be observed in flume experiments, but they cannot be interpreted in the rock record. An understanding of flow transformation would be useful in petroleum geology because Newtonian turbidity currents are more likely to spread out laterally than plastic debris flows in channel-mouth environments (see Chapter 12). Kuenen (195 l) proposed downslope transformations of slumps to mud flows. Subsequently other researchers have proposed transformations of slumping and debris flows into turbidity currents (Dott, 1963; Hampton, 1972). Phillips and Davies (1991, p. 109) noted, '... although a flow may start as a viscous plastic material it may subsequently develop grain-dispersive characteristics. Then, as shear rates are reduced, say, by a reduction in bed slope or by jamming of coarse grains in the channel the flow may once again exhibitplastic-viscoplastic behavior.' In other words, a debris flow may transform into a grain flow, and then revert back to a debris flow. Similarly, transformation of a grain flow into a turbidity current and back to a grain flow during last stages of deposition has been suggested by Middleton (1970). Fisher (1983) proposed four types of transformations for sediment-gravity flows: (1) body transformation; (2) gravity transformation; (3) surface transformation; and (4) elutriation transformation. In experiments on stratified flows, the basal laminar flow was initially fully turbulent, but during the depositional stage the turbulent flow was transformed into a quasi-plastic laminar flow (Postma et al., 1988). This is called gravity-flow transformation. Experimental studies have also shown that plastic debris flows generated Newtonian turbidity currents via surface-flow transformation (Hampton, 1972). If a turbidity current were to generate a basal high-concentration laminar layer due to gravity-flow transformation (Fig. 7.8), Postma et al. (1988) would classify this type a 'high-density turbidity current.' This is because the basal high-concentration layer is derived from and driven by over-tiding turbidity currents. Similarly, if a laminar debris flow were to generate an upper turbulent cloud (i.e., turbidity current) due to surface-flow transformation (Fig. 7.8), one should classify this type a 'low-density debris flow,' following the logic of Postma et al. Such classifications of sediment-gravity flows based on flow transformation are not meaningful. This is because flow transformations cannot be established without knowing: (1) initial flow behavior; (2) transport mechanisms; and (3) final flow behavior. There are, however,

257

Process-related problems Turbidity Current

High - Density Turbidity Current

Debris Flow

Low - Density Debris Flow

r--->

[,.3 ~J

Turbulent Flow

~-~--~

Laminar Flow

Fig. 7.8. Classification of flows based on flow transformations. Top: Gravity-flow transformation of a density-unified flow (turbidity current) into a density-stratified flow (high-density turbidity current). Open arrow shows direction of transport. Based on Postma et al. (1988). Bottom: Surface-flow transformation of a density-unified flow (debris flow) into a densitystratified flow (low-density debris flow). Open arrows show direction of transport. See text for details. (After Shanmugam (2000a). Reproduced with permission from Elsevier.)

no established criteria for recognizing initial flow behavior and transport mechanisms in the depositional record (Dott, 1963; Walton, 1967; Middleton and Hampton, 1973; Carter, 1975; Stanley et al., 1978; Lowe, 1982; Postma, 1986; Middleton, 1993; Shanmugam, 1996a). Many of us use this universal constraint, which is the absence of evidence for transport mechanism in the deposit, as a license to assume that all deep-water sands must have been transported by turbidity currents but that they underwent late-stage plastic transformation so as to resemble debris-flow deposits. If we continue to follow such an assumption-based (i.e., model-driven) interpretation, then there is no need to examine deep-water sands to understand their depositional origin; we can simply assume that all deep-water sands are turbidites. Disappointingly, this is the approach that many pursue in the geologic community (e.g., Hiscott et al., 1997). As a result, the geologic literature is saturated with examples of 'turbidites,' irrespective of whether these sediments were transported and deposited by turbidity currents or by some other processes. During surface-flow transformation, it is common for debris flows to become turbidity currents downslope (Fig. 7.9A). In this scenario, outrunning turbidity currents may initially cut channels that may be filled later by the following sandy debris flows in stratified flows. The upper muddy turbidity currents of stratified

258

Deep-water processes and facies models

Fig. 7.9. (A) Side view showing downdip changes in sediment flows due to surface-flow transformation from a debris flow (left) through stratified flows with a lower debris flow and an upper turbidity current (center) to an outrunning turbidity current cutting channels (right). (B) Front view showing channel filling by lower sandy debris flows and deposition on the levee from upper muddy turbidity currents. In situations like this, the presence of turbidites on the levee does not imply that channels were also filled by turbidity currents.

flows may deposit turbidites on levees (Fig. 7.9B). As a result, the channel-fill deposits would be composed of sandy debrites, and the levee portion would be composed of muddy turbidites. Such a facies association has been observed in the Amazon Fan (see Chapter 6). The presence of turbidites on the levee does not automatically mean that associated channels were also filled by turbidity currents. In discussing the physics of debris flows, Iverson (1997) stated, 'When mass movement occurs, the sediment-water mixtures transform to aft owing, liquid-like state, but eventually they transform back to nearly rigid deposits.' Although such transformations occur during transport, evidence for flow transformations cannot be inferred from the final deposit. We may never resolve this issue of flow transformation because it would be like attempting to establish the previous life history of a human being after reincarnation!

7.6 Conflicting definitions of slurry flows Conventionally, many researchers have considered slurryflows to be debris flows (e.g., Carter, 1975; Mutti et al., 1978; Stanley et al., 1978; Hiscott and Middleton,

259

Process-related problems

1979; Pierson and Costa, 1987). Lowe and Guy (2000), however, equated slurry flows with high-density turbidity currents. Does this mean that slurry flows, debris flows, and high-density turbidity currents are one and the same process? According to Lowe and Guy, deposits of slurry flows contain massive sandstone, sheared sandstone masses in a mud-rich sandstone matrix, subvertical fabric and restructuring by water escape, deformed blobs of sandstone in low-mud sandstone, and floating chunks of sandstone. These features are evidence for deposition from sandy slumps and debris flows (see Chapter 3). In justifying that slurry flows are indeed high-density turbidity currents, Lowe and Guy (2000, p. 65) wrote, 'These cohesion dominated sublayers are analogous in many ways to friction-dominated traction carpets described previously from turbidity currents (Hiscott and Middleton, 1979; Lowe, 1982) and can be termed cohesive traction carpets.' Traction carpets generally develop in mud-free or mud-poor basal granular layers in gravity currents due to dispersive pressure caused by grain collisions. In slurry flows, however, high mud content should greatly reduce the chances of collision between grains and diminish the development of traction carpets. Lowe and Guy did not explain this mechanical paradox. They claim that slurry flows had undergone a number of flow transformations; however, they did not present physical evidence of transformation. This is because all that can be inferred from the depositional record is what happened during the final moments of deposition. Lowe and Guy proposed a sequence of structures for slurry-flow deposits (M1, M2, M3, M4, Ms, M6, and My). More importantly, they suggested that these slurry-flow divisions are comparable to the vertical sequence of fine-grained turbidites (T~, Tb, Tc, Ta, and Te) proposed by Bouma (1962), and to the vertical sequence of coarse-grained turbidites or high-density turbidites (R~, R2, R3, S1, $2, and $3) proposed by Lowe (1982). A comparative analogy of these three facies models results in the following: M1 -- el M2 and M3 = $2 M4 =

S3 =

Ta

M5 = Tb, Tc, Td, and Te M6 and M7--- Post-depositional structures that have no equivalents with structures in either the model of Bouma (1962) or that of Lowe (1982). By comparing the vertical sequence of slurry-flow deposits with the Bouma Sequence, Lowe and Guy implied that the slurry-flow sequence represents a single depositional event. However, slurry-flow beds are amalgamated units that

260

Deep-water processes and facies models

represent multiple, random, depositional events (Lowe and Guy, 2000, see caption of their Fig. 24, p. 61). Thus the proposed vertical sequence of slurry-flow deposits is an attempt to manufacture an artificial order from natural chaos. 7.7 C o n f l i c t i n g origins of flute s t r u c t u r e s

Flute structures that occur as sole marks of deep-water sands were interpreted as turbidites (Hiscott and Middleton, 1979). The assumption being that the head of a turbidity current was turbulent and that turbulence created the scour. Another assumption is that the scour was subsequently filled by the body of the same turbidity current. Care must be exercised in making these assumptions because alternative explanations are possible. Flutes simply suggest that a turbulent state of flow was responsible for creating the scour. Flutes do not imply that the sand that fills a scour surface was deposited by the same turbulent flow that created the scour surface (Sanders, 1965, p. 209). Scour surfaces or depressions on the sea floor, for example, can be created initially by turbulent flows and filled later by laminar sandy debris flows or by other processes (Fig. 7.10). Modem unfilled submarine channels and canyons are a testimony to the fact that the processes that created these erosional features in the past are probably not the same processes that may fill them in the future. Furthermore, scour surfaces can also be created by processes other

Fig. 7.10. Conceptual diagram showing the complex origin of flutes in deep-water sandstone. (A) Smooth sea floor. (B) Formation of depression (flute) by erosion of the sea floor by turbulent flows. (C) Subsequent filling of flute depression by laminar sandy debris flows. (Concept is after Sanders (1965).)

Process-related problems

261

than turbidity currents, such as bottom currents in deep-water environments (Klein, 1966). In a modem tidal flat at Abu Dhabi, Friedman and Sanders (1974) observed positive-relief 'bedforms' that resemble molds of times. These times were ascribed to sheet flow of an incoming tide. Because times can be formed by processes other than turbidity currents, the routine interpretation of times as evidence for turbidite deposition is incorrect. The origin of deep-water sands should be based on their internal depositional features, not on their erosional basal contacts or sole marks. 7.8 Conflicting definitions of normal grading 7.8.1 Single depositional event

The concept of normal grading for turbidites was first introduced by Kuenen and Migliorini (1950, p. 99; see also Kuenen, 1967, p. 212). According to this concept, a turbidite represents a single depositional event by a waning turbidity current. Many researchers have adopted this concept (e.g., Bouma, 1962; Harms and Fahnestock, 1965, p. 109; Sanders, 1965; Middleton, 1967). The link between normal grading and its deposition from a single turbidity current event is the single most important concept in process sedimentology of turbidites. However, Lowe and Guy (2000) applied the concept of normal grading to an amalgamated unit (i.e., their type II slurry) deposited by multiple depositional events of slurry flows. Such an application undermines the sedimentological meaning of normal grading. Mulder et al. (2001) also misused the term 'normal grading' for a unit composed of multiple sand and mud layers deposited from multiple events with alternating intervals of normal and inverse grading. Furthermore, Mulder et al. (2001) included four subintervals of inverse grading as part of normal grading. Clearly, the term normal grading has lost its original process-sedimentologic meaning for a single depositional event (Shanmugam, 2002b). As a result, one can manufacture a normal grading in any depositional package, composed of multiple depositional events, by selectively designating a coarse bottom unit and a fine top unit to achieve the desired outcome. This is one of the common underpinnings of the turbidite controversy. 7.8.2 Simple normal grading

Deposition from turbidity currents commonly occurs through sediment fallout from suspension (Kuenen and Migliorini, 1950; Dott, 1963). In truly turbulent Newtonian flows, coarse and fine-grained particles tend to settle separately during deposition depending on their fall velocities. This causes deposits of turbidity currents to be characterized by simple normal grading (i.e., upward decline in grain size) with gradational upper contacts.

262

Deep-water processes and facies models

Following Kuenen (1953), it is common practice to interpret an entire unit as a turbidite even if grading is restricted to the uppermost portion of the unit. For example, if a two meter thick massive sand has a two centimeter thick normally graded top, only the two centimeter graded top should be interpreted as a turbidite. The origin of the underlying sand requires independent evaluation using its own features. In other words, the two centimeter graded top does not reveal anything about the depositional origin of the underlying massive sand. There are examples in which thick massive sands with rafted ungraded clasts (i.e., a debris-flow origin) have thin graded tops (i.e., a turbidity current origin). 7.8.3 Description of normal grading

To interpret a normally graded unit as a turbidite, one must describe the graded unit with precision. One common pitfall involves description of a massive sand interval as a graded sand. This is achieved by simply connecting the base of a sand unit with the base of an overlying mud unit (Fig. 7.11). Such a practice invariably results in a normal grading. In describing rocks, both megascopic and microscopic details should be documented. Some researchers tend to emphasize primarily microscopic variations

Fig. 7.11. Diagram showing how a massive sand unit, without grading, can be made into a normally graded unit simply by connecting the base of the sand unit with the base of the overlying mud unit.

Process-related problems

263

Fig. 7.12. Diagram showing inverse grading and normal grading of plastic debris flows. Note close sampling intervals for microscopic size analysis may not reveal megascopic features (floating granules and clasts) that are equally important for interpreting depositional processes.

in grain size. Depending on where samples for microscopic grain-size analysis are taken, one can easily overlook the importance of megascopic floating quartz granules and floating clasts (Fig. 7.12). Because both inverse and normal gradings are present in debrites (Fig. 7.12), all details are vital. 7.8.4 Grading in debrites

Although normal grading is the standard criterion for recognizing turbidites, plastic debris flows can also develop normal grading (Vallance and Scott, 1997). An example of a normally graded debrite in the northwestern continental slope of the Gulf of Mexico has been reported by Tripsanas et al. (2003). These normally graded debrite units are less than 1 m thick and consisted of mud matrix with mud clasts (1-5 mm in diameter). Because debrites can produce normal, massive, or inverse grading with or without clasts (Fig. 7.13), recognition of normally graded units without clasts as debrites will always be a challenge. In such cases, X-radiography may be useful for recognizing subtle amalgamation surfaces, buried clasts, and contorted layers. 7.9 Problematic origin of traction structures

Perhaps, the most problematic issue is the use of traction bed forms (i.e., plane bed, ripple, dune, upper flow regime plane bed, and antidune), observed in flume experiments (Simons et al., 1965; see also Southard, 1975), as analogs for the

264

Deep-water processes and facies models

Fig. 7.13. (A) Diagram showing a plastic debris flow with clasts can develop units of normal grading, massive (no grading), and inverse grading with clasts. (B) Diagram showing a plastic debris flow without clasts can also develop units of normal grading, massive (no grading), and inverse grading without clasts. A normally graded debrite unit without clasts may resemble a turbidite.

five divisions of the Bouma Sequence deposited by turbidity currents (see Harms and Fahnestock, 1965; Walker, 1965). Although this analogy is deeply embedded in our psyche (Fig. 7.14), it is founded on unsound hydrodynamic principles. Simons et al. (1965, p. 35) conducted flume experiments in unidirectional currents under equilibrium flow conditions that developed traction structures. Simons et al. (1965, p. 50) cautioned aptly that traction structures observed in their flume experiments are meaningful only for structures developed in subaerial alluvial channels. Walker (1965, p. 22-23) also cautioned that '... the flume experiments were conducted under conditions of non-deposition, whereas many of the sedimentary structures of turbidites are formed under conditions of net deposition.' The time required to establish hydrodynamic equilibrium is greater than the time required for sedimentation (Allen, 1973). Natural turbidity currents are waning flows, and waning flows may never attain equilibrium (Allen, 1973). In most natural flows, changes in bed configurations tend to lag behind changes in flow conditions, and there have been almost no flume experiments on disequilibrium bed configurations (Southard, 1975, p. 33). Thus experimental traction structures are problematic analogs for natural turbidite deposits. The interpretation of deep-water sands with thick divisions of 'parallel lamination' as upper flow regime fiat beds of turbidity currents is problematic. This is because thick upper flat beds have never been generated by experimental

Process-related problems

265

Fig. 7.14. Comparison of size-velocity diagram for bed forms developed in flume experiments at a flow depth of 20 cm (compiled from many sources, see Southard, 1975) with the five internal divisions (A, B, C, D, and E) of the Bouma Sequence (see Chapter 8). Bouma divisions A, B, C, and D are labeled in the size-velocity diagram for fine-grained sand for discussion purposes. Note that dunes and in-phase waves (antidunes) observed in flume experiments are absent in the Bouma Sequence. Also note that the basal normally graded division of the Bouma Sequence is absent in flume structures. Published size-velocity diagram of experimental structures by Southard (1975) is flipped vertically in order to make an easy comparison with the Bouma Sequence. (After Shanmugam (2000a). Reproduced with permission from Elsevier.)

turbidity currents. On the other hand, experimental studies of sandy debris flows have generated horizontal layers that mimic 'parallel lamination' (Fig. 3.24A). The development of pervasive lamination has also been ascribed to sediment shear during freezing of mass flows (Stauffer, 1967). Furthermore, deep bottom currents are traction currents and are quite capable of generating parallel lamination. There is no reason to assume that all laminated beds are turbidites (Murphy and Schlanger, 1962, p. 471). Based on their experimental studies, Leclair and Arnott (2005, p. 4) concluded that, '... the debate on the upward change from massive (Ta) to parallel laminated (Tb) sand in a Bouma-type turbidite remains unresolved.' In bed forms produced in flume experiments of alluvial channels, dunes are an integral part (Simons et al., 1965), whereas cross beds (i.e., internal structures of dune bed forms) are absent in the Bouma Sequence (Fig. 7.14). In fact,

266

Deep-water processes and facies models

Kuenen (1964, p. 16) used the presence of dune structures in deep-water sandy sequences as evidence against turbidite deposition. Kuenen (1964, p. 25) stated, '... a turbidity current must have carried its load of grains in suspension almost up to the point at which each particle comes to rest.' For this reason, normal grading, which is typical of turbidite beds, is absent in experimental structures of alluvial channels (Fig. 7.14). Large-scale cross bedding (dune bed forms) is generally absent in Bouma-type turbidites (Picketing et al., 1989). Their absence is ascribed to various causes, such as flows being too rapid (Walker, 1965), flows being too thin (Walker, 1965), flows being too fine-grained (Walton, 1967), or flows with Froude number greater than 0.35-0.40 (Hsfi, 1989). Although cross bedding has never been generated by turbidity currents in laboratory experiments, turbidity currents of varying densities have been proposed to explain the origin of cross beds. They are" (a) high-density turbidity currents (Lowe, 1982); (b) moderate- to low-density turbidity currents (Piper, 1970; Winn and Dott, 1979); and (c) low-density turbidity currents (Martinsen, 1994). In attributing deep-water antidune 'cross bedding' to high-concentration turbidity currents (i.e., high-density turbidity currents), Picketing et al. (2001, p. 698) stated, 'The surface morphology and internal structure of the inclined sandy macroforms is inconsistent with well-constrained large-scale antidunes formed in fluvial environments. Although large-scale deep-water bedform fields interpreted as antidunes remain poorly studied, without clear contrary observations it seems reasonable to assume that their architecture would be similar to fluvial examples.' The problems with high-density turbidity currents were discussed earlier. Piper et al. (1988) suggested that deep-sea gravel waves in the Grand Banks area are products of bed load transport by turbidity currents, analogous to dune bed forms in subaerial rivers that involve traction processes. The implication is that these gravel waves are composed of cross bedding; however, no core information is available to prove the presence of cross bedding. Hsfi (1989) proposed an alternative origin for gravel waves by debris avalanches. Until we establish a genetic link between turbidity currents and cross bedding either by direct observation in deep-water environments or by experimental studies in the laboratory, the origin of cross bedding by turbidity currents will remain controversial. 7.10 Problematic origin of mud waves

Large-scale features (10-80 m in height), such as 'migrating mud waves' or 'abyssal bedforms,' have been reported in the deep sea ( Flood, 1988; Klaus and Ledbetter, 1988). However, these deep-water features should not be equated with dune bed forms in rivers that create cross bedding due to bed load transport of granular material, which must be larger than 125 microns in grain size (i.e., fine sand).

Process-related problems

267

Deep-sea migrating waves are composed primarily of silt and clay, and therefore they do not have the necessary sand grade (i.e., granular material) to generate cross bedding. Mud waves are ascribed to sculpting of muddy sea floor by deep bottom currents, such as the Antarctic Bottom Water (AABW) in the Argentine Basin (Klaus and Ledbetter, 1988). The precise mechanism of mud wave origin is unclear. Giant sediment waves (5 m in height) on the continental margin off Nice (southern France) that are composed of sand and boulders were ascribed to deposition by 'sediment flows' (Malinverno et al., 1988). These sediment flows have been considered to be a combination of both debris flow and turbidity current (Malinverno et al., 1988). However, it is unclear how these sediment waves were deposited by two rheologically different sediment flows composed of plastic debris flows and Newtonian turbidity currents.

7.11 Problematic subaerial analogs A troubling practice is to compare subaqueous turbidity currents with subaerial river currents (Chikita, 1989). This comparison is incorrect for many reasons. River currents and turbidity currents are fundamentally different, although both are turbulent in state. River currents are generally low in suspended sediment (1-5% by volume; Galay, 1987), whereas turbidity currents are relatively high in suspended sediment (1-23% by volume; Middleton, 1967, 1993), although both currents are considered to be Newtonian in rheology. River currents are fluidgravity flows, whereas turbidity currents are sediment-gravity flows. A common perception is that high-density turbidity currents in subaqueous environments are analogous to hyperconcentrated flows in subaerial environments (Fig. 7.15). According to Pierson and Costa (1987), hyperconcentrated flows are plastic flows, and thus high-density turbidity currents cannot be true turbidity currents. In China (Qian et al., 1980), the term hyperconcentrated flow is used for two distinctly different types of flows: (1) Newtonian fluids characterized by a low sediment concentration and a turbulent state in which coarse and fine particles settle separately (i.e., settling); and (2) Bingham (i.e., non-Newtonian) fluids characterized by a high sediment concentration and a laminar state in which coarse and fine particles are deposited together (i.e., freezing). Analogous to application of the term hyperconcentrated flow for both turbulent and laminar flows in subaerial environments (Qian et al., 1980), the term high-density turbidity current is applied to both turbulent and laminar flows in subaqueous environments (Postma et al., 1988). High-density turbidity currents are commonly perceived to occupy an intermediate position between low-density turbidity currents and debris flows (Fig. 7.15). The problem is that both hyperconcentrated flows and debris flows are considered to be non-Newtonian fluids that exhibit plastic behavior (Fig. 7.15). Thus the use of subaerial processes as analogs for subaqueous high-density turbidity currents is problematic.

268

Deep-water processes and facies models

Onset of Yield Strength ~,

Subaerial Fluid Behavior Flow

l lNewtonian;.,,,,,, ,, :""1 Ill, .... i Ii i!'""'""" iiiiii

1.,1

i,uqul O I,

,

Non-Newton,an

"

[I Ii I . . . . . . . . . .

ll,ll,I

iI.

Plastic --

- . . . . . . . . . . . . . . . . . . .

I"iN6imai "S}re~r~'i:low"" I::"H~)perc6n6ehtr'aiea " ~ ~ ~ o ~" 9 9 . .']..'. :-" Stream Flow'.'.; .'.'~ .'1 o ."o ~Deboris.Fl~176 (Pierson and Costa, 1987)

Subaqueous Flow

l'i" "":'l'~-....., ~ "'Turhlrglh' .

.

.

.

i

" 1;." " Turl:)ldity :'High'D'e-I;lsity-'"'-'" ~ . " ." ro ~ C"rrent'"

9

-

"

"

"

,

i

"Debris ~ . .Flow; . .

o

o

o

o

O

o

- :'1

(Common Perception)

Subaerial

i ,

II[I I Ne~vtonian I I i I ' il- ..........

Bingham - i:_.=.! @~~

.... __--- 7 7 -_ .... .I "" :" """"" "'~" i': "."" """ "" filtiJercofice'ntratecl'.lBIrw. "'"" :. . i1., !t., I., I'j '.'l' ul-' ,17 _ - . . - .

Flow

-

"

-

9

l

9

"

"

"

"

*

".

~

i

l"

!

"1

9

o.

9

t

~

"

~

"

i

i"

!i't

Fig. 7.15. Schematic diagram showing inconsistencies when comparing subaerial 'hyperconcentrated flow' with subaqueous 'high-density turbidity current" (After Shanmugam (2000a). Reproduced with permission from Elsevier.)

7.12 Problematic origin of sinuous forms

Sinuous channel forms have been observed in both modem and ancient deep-water systems. Their origin in deep-water environments has not yet been resolved. This is because sinuous forms have been associated with seven contrasting processes: (1) fluvial helical flows; (2) turbidity currents; (3) debris flows; (4) volcanic lava flows; (5) glacial melting; (6) droplets on a glass plate; and (7) faulting. 7.12.1 Fluvial helical flows

In a fluvial channel, a meander (i.e., a stream with sinuosity >1.5) represents a curve or a bend. The outside portion of a meander, the concave feature known as a cut bank, is created by erosion. The inner portion of a meander, the convex feature known as a point bar, is created by deposition (i.e., lateral accretion). The point bar is formed when sediments drop from the bed load (traction) as the water velocity of helical flow at the meander slows upslope. In this setting, as the flow moves upslope over the inclined point-bar surface, bed shear stress decreases upward. As a consequence, larger bedforms with coarser grains are deposited near the bottom and finer grains with smaller bedforms are deposited near the top (Allen, 1964). Although the link between helical flows and point-bar deposition has been established, the precise mechanism by which a meandering channel develops in fluvial systems is still poorly understood.

Process-relatedproblems

269

There were attempts to explain deep-water meanders directly by fluvial mechanisms. This was accomplished by: (1) forming fluvial meanders in subaerial environments, and (2) later creating a deep-water environment at the site of fluvial meanders by a rapid rise in sea level. Such drastic changes in sea level are geologically unrealistic to explain deep-water meandering channels at a water depth of 3000 m (e.g., Amazon Fan, Atlantic Ocean). 7.12.2 Turbidity currents

The origin of sinuous channel forms in deep-water lower Miocene of offshore Angola has been explained by turbidity currents (Abreu et al., 2003). They called these turbidites as Lateral Accretion Packages (LAPs), and stated (p. 645), 'The sinuous and erosionally confined Green Channel Complex, however, displays strong lateral and downdip migration of channels, accretion deposits at the inner portion of every channel bend (LAPs) and frequent cut-off meanders (about 20% of the meanders in the studied area). The apparent similarity of these deepwater deposits (LAPs) to fluvial point-bar deposits suggests that the controls on their formation may be similar. Therefore, the depositional model proposed for these deepwater LAPs is that the accretion surfaces would be formed by relatively continuous and gradual lateral sweep of channel bends by systematic erosion of the outer banks and deposition along inner banks .... This is similar to the classic fluvial point-bar model (Galloway and Hobday, 1983).'

In this case, the implication is that deep-water turbidity currents behaved like fluvial helical flows and developed deep-water deposits of Angola. If true, then the Angolan deposits should exhibit the following depositional features that are characteristic of point bars: (1) lateral accretion package with a basal lag composed of bed load gravel; (2) imbrications of pebbles and cobbles; (3) large dune bedforms (cross bedding) at the lower part of the point bar; (4) ripples and plane beds at the upper part of the bar; and (5) mud drape by vertical accretion. However, none of these features has been documented by Abreu et al. in cores from Angola. Interestingly, Abreu et al. (2003, their Fig. 23) used the Pennsylvanian Jackfork Group, which has served as the epicenter ofturbidite controversy (see Chapter 9), as an outcrop analog for the Angolan sinuous channels. In this outcrop analogy, three lithofacies are used: (1) thick bedded Ta; (2) thin bedded Ta; and (3) mudclast conglomerate. The origin of Ta division in the Bouma Sequence has been attributed to eight different processes (see Chapter 8). None of the eight processes can be related to helical flows in sinuous channels. More importantly, the Jackfork Group has been interpreted to be sandy debrites (see Chapter 9). Debris flows with plastic rheology cannot behave as helical flows in meander bends. Unlike prolonged fluvial currents that develop sinuous channel morphology on land, turbidity currents are episodic. Thus it is difficult to envision how such

270

Deep-water processes and facies models

episodic events could develop the highly meandering channel forms observed in the deep sea. Abreu et al. used plan-view trends seen on seismic time slices and crosssectional geometry seen on seismic profiles for developing their depositional models. These geometry-based models have to be ultimately validated by processbased models developed from cores. Unfortunately, no 'ground truth' has been presented from Angolan cores. Our tendency to equate fluvial point-bar deposition by traction processes with turbidite deposition in deep-water sinuous channels by suspension processes is hydrodynamically not meaningful. Peakall et al. (2000) cautioned that the use of fluvial analogy for submarine channels is superficial. Importantly, experimental studies have failed to generate point bars in subaqueous meanders (M&ivier et al., 2005). 7.12.3 Debris flows

Sinuous channels of the Amazon Fan are dominated by deposits of sandy debris flows and slumps (see Chapter 6). The close association between sinuous canyons/channels and mass-transport deposits is also evident in modem slopes off the Los Angeles Margin, Lake Tahoe, the Monterey Fan, and the Mississippi Fan (see Chapter 6). The genetic link between sinuous channels and mass-flow deposits is unclear. In subaerial channels with sharp bends, the cross-section of channels after debris flows had passed through showed occurrence of spill-over deposits on both sides of the channel (Johnson, 1984). Spill-over deposits have also been reported from submarine debris-flow chutes (Prior and Coleman, 1984). The geometry of the spill-over deposits of debris-flow chutes may mimic the popular gull-wing geometry of channel-levee complexes associated with sinuous channels (see Chapter 6). There are, however, no established criteria in seismic profiles to distinguish the origin of a 'gull-wing' geometry by turbidity current versus by debris flow. Future experimental research should focus on the development of sinuous channels and channel-levee complexes by debris flows. 7.12.4 Volcanic lava flows

Volcanic lava flows from Pu'u 'O' o cone in the Kilauea Volcano, Hawaii have generated channel forms ranging from braided to sinuous patterns (Fig. 7.16). In fact, the fluvial term levee has been applied to describe the behavior of lava flows (Sparks et al., 1976). There are clear plan-view patterns of lava flows that can be labeled as meandering channel, braided channel, levee, splay, and lobe. However, these terms do not reveal anything about depositional processes. Although lava flows represent a wide range of Reynolds Number, such flows are commonly plastic in rheology (Johnson, 1970; Griffiths, 2000). Similar to sinuous lava flows,

Process-relatedproblems

271

Fig. 7.16. Plastic lava flows showing sinuous and braided channel forms, Pu'u 'O' o cone, Kilauea Volcano, June 2, 1986. Credit: J. D. Griggs, USGS, Hawaiian Volcano Observatory. Uniform Resource Locator (URL): http://hvo.wr.usgs.gov/hazards/dds24167_photocaption. html (accessed February, 2005).

submarine debris flows with plastic rheology may also create sinuous patterns that may mimic sinuous channel forms in seismic amplitude slices. 7.12.5 Glacial melting

Melting glacier has developed meandering canyons in the Kennicott Glacier, Alaska (Mitchell, 2003, p. 75). Glacial meltwaters are hyperconcentrated flows and they often flow under pressure. The exact mechanism of these glacial meanders is unclear. 7.12.6 Droplets on a glass plate Dewdrops from condensation of atmospheric moisture commonly develop meander patterns on a car windshield on a cold morning. Janosi and Horvath (1989) studied the phenomenon of meandering droplets on a glass plate. A water droplet begins to run down an inclined glass plate if the mass exceeds a static critical weight.

272

Deep-water processes and facies models

This depends mainly on interfacial tensions and on the slope of the glass. The route of the meandering droplet is determined by impurities on the surface as well as in the droplet itself. In spite of many theoretical and experimental studies, precise mechanism of this phenomenon is still poorly understood. 7.12.7 Faulting

A structural origin, controlled by fault motions, of tight meander in the Monterey Canyon was discussed earlier (see Chapter 6). In summary, at the time of this writing (May 2005) there is no known mechanism that can explain the origin of sinuous channel forms in deep-water environments. Thus equating deep-water meanders seen on seismic amplitude slices with fluvial point bars is imprudent. The dangers of using seismic attributes in geologic interpretation have been known (Sheline, 2005). Since the recognition of a tight meander of the Monterey Canyon (Offshore California) by Shepard (1966), the number of papers published on deep-water meanders has been phenomenal, but the depth of our understanding of their origin has not been proportional. An unfortunate growing trend has been to reinvent concepts of the 1960s and to repackage them under erudite nomenclature. 7.13 Problematic hyperpycnal flows

The concept of hyperpycnal flows is popular because of the misperception that it can be used for explaining some complicated features of deep-water successions such as inverse grading (Mulder et al., 2002, 2003). Because inverse grading is an important feature of deep-water sands (see Chapter 8), the concept of hyperpycnal flows requires a critical evaluation. In advocating a rational theory for delta formation, based on the concepts of Forel (1892), Bates (1953) suggested three major flow types: (1) hypopycnal flow for floating river water that has lower density than basin water (Fig. 7.17A); (2) homopycnalflow for mixing river water that has equal density as basin water (Fig. 7.17B); and (3) hyperpycnalflow for sinking fiver water that has higher density than basin water (Fig. 7.17C). Although hyperpycnal flows were designed to describe river-dominated deltaic systems, they have been compared with turbidity currents. For example: (1) In their study of the Huanghe (Yellow River) delta front, Wright et al. (1986, p. 99) stated, '... high density hyperpycnal plume are most like the classical turbidity currents .... ' (2) Wright et al. (1986, p. 98) suggested that '... the hyperpycnal plumes are sustained (as opposed to episodic) turbulent suspension currents, or turbidity currents.'

Process-relatedproblems

273

Fig. 7.17. Schematic diagrams showing three types of density variations in river water in deltaic environments. (A) Hypopycnalflow in which density of river water is less than density of basin water. (B) Homopycnalflow in which density of river water is equal to density of basin water. (C) Hyperpycnalflow in which density of river water is greater than density of basin water. (Based on concepts of Bates (1953).)

(3) Mulder et al. (2002, 2003) distinguished a slow-moving hyperpycnal turbidity current as a special type offlood-generated turbidity current that is different from slide-generated surge-type turbidity currents. More importantly, Mulder et al. (2003) equated hyperpycnal turbidity currents and surge-type turbidity currents with low-density and high-density turbidity currents of Lowe (1982), respectively. Mulder et al. (2003) expanded the applicability of the concept of hyperpycnal turbidity currents from shallow-water (deltaic and continental shelf) to deepwater (continental slope and abyssal plain) environments. The application of the concept of hyperpycnal flows to deep-water systems is problematic for the following reasons: (1) There is no documented evidence for the direct transportation of sediment from river mouths to abyssal plains by hyperpycnal flows in modem oceans. During periods of lowstands, hyperpycnal flows may play a role in shelfedge deltaic settings that may promote sediment transport into deep-water environments. It is difficult to envision the long-distance travel of such

274

(2)

(3)

(4)

(5)

(6)

Deep-water processes and facies models

low-velocity hyperpycnal flows from river mouths to abyssal depths during periods of highstands on the Atlantic-type passive continental margins. Bates (1953) introduced the concept ofhyperpycnal flow for fiver-dominated deltaic processes, whereas Bell (1942) introduced the concept of underflow for deep-water density currents. As a synonym for Bell's (1942) underflow, Mulder and Alexander (2001) used the term hyperpycnal turbidity current. Dasgupta (2003, p. 270) critiqued the hyperpycnal flow concept of Mulder and Alexander as follows: '... use o f the term hyperpycnal to qualify a type of turbidity current is not only superfluous, but to some extent, also creates ambiguity in understanding turbidity current.' Prior and Bomhold (1990) proposed that Holocene fan deltas originated by a combination of processes, namely: (1) debris avalanching; (2) inertia flows (i.e., hyperpycnal flows); and (3) turbidity currents. In explaining inertia flows, Prior and Bomhold (1990, p. 80-81) stated, 'The near-bed or 'movingbed'concentrations of sediment are largely unaffected by seawater density and thus begin moving across the subaqueous slope as hyperpycnal flow .... Gravel~coarse sand transport downslope over the fan appears to be achieved by high-density, pseudo-laminar 'inertia flows'described experimentally by Postma et al. (1988). These authors illustrated a mechanism by which coarse particles move over the bottom by a combination o f dispersive pressure, hindered settling and enhanced buoyant lift.' Clearly, these hyperpycnal flows are laminar flows, not turbulent turbidity currents. Mulder et al. (2003, their Table 1) classified certain major rivers such as the Mississippi, Amazon, and Zaire as clean rivers, rivers that cannot produce hyperpycnal flows. The problem with their classification is two fold: (a) Mulder and Syvitski (1995) originally classified the Zaire as a clean river that cannot produce hyperpycnal flows, but they now believe that the Zaire is a dirty fiver that can produce hyperpycnal flows (Mulder et al., 2003). If so, the classification of clean versus dirty rivers is not based on objective criteria. (b) The Mississippi and the Amazon Rivers have developed two of the largest submarine fans in the world. If these huge fans were not qualified to receive sediment from hyperpycnal flows, in the geologic record how would we know which deep-water fans received sediment from hyperpycnal turbidity currents and which ones received sediment from slide-generated surge-type turbidity currents? Mutti et al. (2003b) expressed doubts about sand deposition from hyperpycnal flows and stated, '... even in the case o f the Yellow river, the only modern large river that generates semi-permanent underflows at its mouth (Van Gelder et al., 1994), there & no evidence o f substantial sand deposition in the delta front region.' The synonymous use of deltaic hyperpycnal flows for deep-water underflows is problematic because underflows can be generated solely by differences in temperature and/or salinity (e.g., Gould, 1951). As a result, geostrophic

Process-related problems

275

thermohaline currents, which are generated in the oceans solely by differences in temperature and salinity (see Chapter 4), could be considered to be underflows. Thermohaline currents, popularly known as contour currents, are density currents. As such these sediment-starved, bottom currents are neither sediment-gravity flows nor turbidity currents. Although all turbidity currents are density currents, but not all density currents are turbidity currents! (7) Harms (1974) advocated a concept of deep-water density current, generated by increased salinity and lowered temperature, to explain traction structures in the Permian Brushy Canyon Formation, West Texas. Harms pointed out that these density currents were not turbulent turbidity currents. The term hyperpycnal flow should be restricted to just deltaic processes. Otherwise, the danger exists for interpreting deep-water bottom-current deposits (e.g., contourites) as a special type of turbidites, known as hyperpycnite (e.g., Mulder et al., 2002, p. 119). For these reasons, Kuenen (see Sanders, 1965, p. 217) emphasized that any definition of turbidity currents should exclude muddy rivers (i.e., hyperpycnal flows) and normal marine currents (i.e., contour currents) from consideration. Or else, we might as well simply abolish specific process terms, such as turbidity currents and contour currents and revert back to the days of density currents of the 1930s.

7.14 Conflicting origins of massive sands Deep-water massive sands have been defined as '... very thick (>1 m) sand beds or units that are devoid o f primary sedimentary structures and that occur in association with other deep-water sediments ...' (Stow and Johansson, 2000, p. 145).

Other researchers consider deep-water massive sands as beds less than one meter in thickness (Rodriguez and Anderson, 2004). Although there is general consensus that massive sands are structureless, their origin is controversial (Parize et al., 1999; Stow et al., 1999). A 400-ft (122 m) thick massive sand unit in the Gryphon Field (U.K. North Sea) contains floating mudstone clasts and internal shear plane (Fig. 7.18). This massive sand has been interpreted as deposits of sandy debris flows and slumps (Shanmugam et al., 1995a). The reservoir sand in the Gryphon Field has also been interpreted as sand injections (Purvis et al., 2002). The controversy over the origin of deep-water massive sands is of economic importance because many deep-water massive sands are major petroleum reservoirs. Selected examples of deep-water massive sands associated with petroleum exploration, discovery, and production are (Stow and Johannson, 2000)" Paleogene fields in the North Sea: (1) Alba* (*means that I have described cores from these examples) (2) Gryphon*

!

Fig. 7.18. An example of deep-water massive sand. (A) Well developed blocky wireline-log motif of a 400-ft (122 m) thick sand, Lower Eocene, Gryphon Field, North Sea. (B) Depth-tied sedimentological log showing massive sand with dish structures, floating mudstone clasts and a basal shearing plane. (C) Core photograph showing a floating mudstone clast in massive sand. Arrow shows stratigraphic position of clast. Well: Kerr-McGee 9/18b-7. See Chapter 10 for a case study of the Gryphon Field.

Process-related problems

(3) (4) (5) (6) (7) (8) (9)

277

Balder* Frigg* Heimdal Forties Montrose Arbroath Andrew*

Mesozoic Fields in the North Sea: (1) (2) (3) (4) (5) (6) (7)

Agat* Kopevic Galley Claymore Scapa Sonjefjord Brae-Miller complex Offshore Brazil:

(1) Marlim* (2) Carapeba (3) Macae West Africa: (1) (2) (3) (4)

Baudroie Baliste Zafiro* Opalo*

West of Shetlands (1) Foinaven* (2) Schiehallion Gulf of Mexico* Seven processes have been proposed for the origin of deep-water massive sands: (1) Low-density turbidity currents: Massive sands in the Annot Sandstone (SE France) were ascribed to the basal part of the turbidite facies model (Bouma, 1962). These massive beds were interpreted as deposits of turbidity currents (i.e., classical, low-density, turbidity currents). The problems with the turbidite facies model and its controversial origin are discussed in Chapter 8.

278

Deep-water processes and facies models

(2) High-density turbidity currents: On the basis of experiments, Middleton (1967) proposed that massive parts of turbidite beds could result from highconcentration turbidity flows (i.e., high-density turbidity currents). However, Middleton's experimental high-concentration turbidity flows met all the criteria for mass flows (i.e., debris flows), as defined by Dott (1963): (a) these flows were flows of non-Newtonian fluids that exhibited plastic behavior; (b) these flows were high-concentration flows in which the sediment was supported by dispersive pressure; and (c) deposition from these flows occurred by 'freezing.' Therefore, the routine interpretation of deep-water massive sands as high-density turbidites is inappropriate. (3) Sandy debris flows" The origin of massive sands has been attributed to deposition from sandy debris flows (Shanmugam, 1996a). Massive sands have been produced by experimental sandy debris flows (see Chapter 3). (4) Sandy debris flows and high-density turbidity currents" The origin of massive sands has been attributed to both sandy debris flows and high-density turbidity currents (Stow and Johanssen, 2000). In an attempt to resolve the controversy over the origin of massive sandstones by high-density turbidity currents versus sandy debris flows, Baas (2004) presented a mathematical model (TDURE). In this model, deposits of high-density turbidity currents are presumed to develop traction structures in their upper parts. According to Baas, deep-water massive sands in which upper traction structures are absent must be due to post-depositional destruction of traction structures by liquefaction, bioturbation, and/or erosion. Baas (2004, p.309) concluded that 'The lack o f stratification does not necessarily mean that massive sands are generated by 'en masse 'freezing o f a debris flow.' The clear implication here is

that all deep-water massive sands should be presumed to be deposits of highdensity turbidity currents. Developing a mathematical model for a concept that is yet to be defined on flow density is of little value for solving the controversy. In an observational science like physical sedimentology (Allen, 1985a, p. ix), a mathematical model is not a substitute for observations of details from the rocks or for experimental observations. (5) Quasi-steady concentrated density currents" The origin of massive sands has been attributed to quasi-steady concentrated density currents (Mulder and Alexander, 2001). However, Mulder and Alexander (2001, p. 288) acknowledged that 'There are no published records o f such flows, and they are likely to be rare occurrences .... ' It is difficult to evaluate scientifically a flow that has been undocumented. (6) Sand injections" Duranti and Hurst (2004) proposed fluidized sand injection as a mechanism for forming deep-water massive sands. This is a viable option, but distinguishing injected sands from debrite sands in cores is a challenge (see Chapter 5).

Process-relatedproblems

279

(7) Contour currents: A contourite origin for massive sands on shelf and upper slope, offshore Antarctica has been proposed by Rodriguez and Anderson (2004). These massive sands are attributed to deposition by Circumpolar Deep Water (CDW). This sandbody is composed of fine- to medium-grained massive sand. Its thickness ranges from 10-100 cm and it forms a large sheet

(3200 km.2). The most likely explanations for the origin of most deep-water massive sands are sandy debris flows and sand injections.

7.15 Conflicting definitions of turbidite systems Different authors have defined the term 'turbidite systems' differently, but none of the definitions makes process sedimentological sense. For example: (1) Mutti (1985, p. 70) defined a turbidite system as '... a body of rocks where channel-fill deposits are replaced by nonchannelized sediments in a downcurrent direction.' Mutti also proposed three types of ancient turbidite systems on the basis of where sand is deposited. In the Type I system, the bulk of the sandstone occurs in detached lobe deposits. In the Type II system, sandstone occurs in attached lobes and channel-lobe transition areas; and in the Type III system, sandstone occurs in channel-fill complexes. (2) Mutti and Normark (1987) classified deep-water systems into five hierarchical orders: (a) A first order Turbidite Complex represents basin-fill scale. It is composed of multiple 'Turbidite Systems' deposited over several millions of years. (b) A second order Turbidite System represents an individual depositional system or a fan. It is deposited over hundreds of thousands of years. (c) A third order Turbidite Stage represents features within a 'Turbidite System'. It formed during tens of thousands of years. (d) Fourth order Turbidite Elements represents lobes and channels, and their facies associations. They are deposited during thousands of years. (e) Fifth order Turbidite Beds represents individual beds, and they are deposited essentially instantaneously. In this classification, all deep-water deposits (e.g., debrites, slumps, and slides) are labeled as turbidites. (3) Bouma et al. (1985) used the term submarine fans for modern deep-water systems, and the term turbidite systems for ancient deep-water systems. (4) Stelting et al. (2000, p. 2) defined the term turbidite systems as deposits of gravity flows. This is confusing because gravity flows (i.e., sedimentgravity flows) are composed of debris flows, grain flows, fluidized sediment flows, and turbidity currents. The term turbidite represents deposits of turbidity currents exclusively (Sanders, 1965). Therefore, to preserve the

280

Deep-water processes and facies models

original meaning of the term turbidite for deposits of tubidity currents, the term turbidite systems should also be used to represent deep-water systems that are dominated by turbidites (Shanmugam, 2001). Without such constraints, there is a danger of classifying deposits of debris flows and slumps under the catchall term turbidite systems.

7.16 I n a d e q u a t e s e i s m i c r e s o l u t i o n

Based on seismic facies and geometries, Vail et al. (1991) classified deep-water systems into basin-floor fans and slope fans in a sequence-stratigraphic framework. In turn, they used these fan models to predict specific depositional facies composed of turbidites (see Chapter 10). The practice of interpreting turbidite facies from seismic data can be misleading and even meaningless. This is because (1) interpretation of turbidite facies requires conventional core or outcrop (centimeter- to decimeter-thick turbidite units cannot be resolved in seismic data (Fig. 7.19); (2) seismically resolvable, thicker (> 30 m), packages are composed commonly of more than one depositional facies; and (3) a single depositional facies can generate more than one seismic geometry (see Chapter 10). Until we systematically calibrate seismic facies with process sedimentology by use of long cores, process interpretation of seismic data using templates of seismic stratigraphy and seismic geomorphology is only an exercise of our imagination with little hydrodynamic basis.

Fig. 7.19. Diagram showing that turbidite beds, commonly ranging in thickness from a few centimeter to a meter, cannot be resolved in standard seismic data.

7.17 S y n o p s i s

In reviewing the literature for this book, it has become abundantly clear that fundamental concepts of turbidity currents were gradually reversed in the 1970s.

Process-related problems

281

The Karma of Turbidity Currents and Their Deposits l llll

1 9 5 0 s and 1 9 6 0 s

1970s

1 9 8 0 s and 1990s

Subaqueous

Subaerial

Turbulent Flow

Laminar Flow

Waning Flow

Waxing Flow

Suspended Load

Bed Load

Newtonian Rheology

Plastic Rheology

Normal Grading

Inverse Grading

Dirty Sand

Clean sand

Fig. 7.20. Diagram illustrating that fundamental concepts of turbidity currents were gradually reversed in the 1970s without scientific basis.

There is no scientific basis for this reversal. It appears as if the turbidite paradigm had passed through a twilight zone (Fig. 7.20). The problem is that on the one extreme, turbidity currents are considered to be turbulent flows, and on the other extreme, they are laminar flows. Under these two extremes (Fig. 7.20), virtually any liquid that moves can be classified as a turbidity currem! During the past 50 years, our purpose has shifted from one of describing and interpreting depositional origin of deep-water rocks to one of justifying their presumed turbidite origin. For no good scientific reason, we have elevated the concept of turbidity current from an ordinary process into an extraordinary phenomenon. That is the karma of turbidity currents!